Will 2.6 require Python for any configuration ? (CML2)
Quoted 3 lines
The main reason to include it, is that that's what it was done in. If you go back and read the archives, ESR goes over why all sorts of other languages wouldn't work as easily.
in such cases the solution is to elaborate, and not to leave things to decay.
Quoted 6 lines
That wasn't my point at all. My point was that if you're somehow transfering the 21mb source .tar.bz2'ed, you can also stand to transport the 4mb of python 2.0.1 source, tar.gz'ed over as well. In other words, having to bring python over any of the methods that Jes mentioned isn't any more painful than the kernel source. It's roughly the size of a couple of vmlinux'es.
i was sarcastic here. actually the fact is that
4MB of tarred sources is more than 10 .c files
doing the same thing 1.5x times faster.
Quoted 5 lines
Have you tried cml2 on your p166? ESR went and did much speed tweaking of the code about 6 months ago it seems like and managed to please some of the people using a low-end pentium. Building a kernel on a 386 isn't approcaching tolerable right now anyhow. Someone pointed out today or yesterday it takes ~10 days.
it is not an excuse to make things even worser.
Quoted 1 line
Python is no harder to maintain then C.
actually i meant that "i hardly can believe that c in such task is harder to maintain than python".
---
cheers,
Samium Gromoff
Quoted 6 lines · Samium Gromoff
Quoted 3 lines
The main reason to include it, is that that's what it was done in. If you go back and read the archives, ESR goes over why all sorts of other languages wouldn't work as easily.
in such cases the solution is to elaborate, and not to leave things to decay.
Well, the flames finally died down and everyone forgot about it again for a while.
Quoted 9 lines
Quoted 6 lines
That wasn't my point at all. My point was that if you're somehow transfering the 21mb source .tar.bz2'ed, you can also stand to transport the 4mb of python 2.0.1 source, tar.gz'ed over as well. In other words, having to bring python over any of the methods that Jes mentioned isn't any more painful than the kernel source. It's roughly the size of a couple of vmlinux'es.
i was sarcastic here. actually the fact is that
4MB of tarred sources is more than 10 .c files
doing the same thing 1.5x times faster.
Where'd you get 1.5x from? :) But, go ahead and re-write the CML2 engine in C. There's been some work done on this. A few people (Al Viro?) even said that if this got into 2.5 they'd do a 'C' version. Many many months ago, if i recall correctly.
Quoted 6 lines
Quoted 5 lines
Have you tried cml2 on your p166? ESR went and did much speed tweaking of the code about 6 months ago it seems like and managed to please some of the people using a low-end pentium. Building a kernel on a 386 isn't approcaching tolerable right now anyhow. Someone pointed out today or yesterday it takes ~10 days.
it is not an excuse to make things even worser.
It shouldn't be a consideration at all. I'd go as far as to say any sort of speed bonus on a 486 is gravy. In other words, how fast something will be on 10 year old hardware shouldn't be important (I'm speaking of 386 with 10 years.)
Signature · 2 lines
Tom Rini (TR1265) http://gate.crashing.org/~trini/